@paoloredaelli and

>>> @g
... def h() -> float:
...     """I'm like f but with a new name and already decorated"""
...     return random.random()
...
>>> f()
0.9379232620733947
>>> g(f)()
10.1676747637114
>>> h()
10.486860853177589

In the example above we are doing g ∘ f. The @ syntax is just syntactic sugar.

This is the core concept, but usually decorators have a more practical applications like in the examples of your quoted article like performance measurement.

@paoloredaelli you can think decorators like mathematics function composition.

For example (a really silly one, but only to get the idea):

>>> import random
>>> import typing as t
>>> def g(f: t.Callable) -> t.Callable:
...     def add_ten():
...         return f() + 10
...
...     return add_ten
...
>>> def f() -> float:
...     """Return a random value in [0.0, 1.0)"""
...     return random.random()
...